Introduction Although diagnostic instability in first-episode psychosis (FEP) is of major concern, little is known about its determinants. Objectives To examine the long-term diagnostic stability of FEP diagnoses, the baseline predictors of diagnostic change to schizophrenia and the timing of diagnostic change. Methods This was a longitudinal and naturalistic study of 243 subjects with FEP who were assessed at baseline and reassessed after a mean follow-up of 21 years. The diagnostic stability of DSM-5 psychotic disorders was examined using prospective and retrospective consistencies, logistic regression was used to establish the predictors of diagnostic change to schizophrenia, and survival analysis was used to compare time to diagnostic change across diagnostic categories. Results The overall diagnostic stability was 47.7%. Schizophrenia and bipolar disorder were the most stable diagnoses, with other categories having low stability. Predictors of diagnostic change to schizophrenia included a family history of schizophrenia, obstetric complications, developmental delay, childhood adversity, poor premorbid functioning in several domains, long duration of untreated continuous psychosis, spontaneous dyskinesia, lack of psychosocial stressors, and poor early treatment response (Table 1). There were no significant differences between specific diagnoses regarding time to diagnostic change. At 10-year follow-up, around 80% of the diagnoses had changed. Table 1. Main baseline predictors of diagnostic change to schizophrenia over the follow-up (univariate logistic regression) No diagnostic change (n=124) † Diagnostic change (n=47) † OR (95% CI) p Family history of schizophrenia spectrum disorders 16 (12.9) 16 (34.0) 3.48 (1.56 – 7.75) 0.002 Obstetric complications, any definite 9 (7.3) 12 (25.5) 4.38 (1.70 – 11.2) 0.002 Developmental delay at year 3, any 30 (24.2) 30 (63.8) 5.52 (2.68 – 11.3) <0.001 Childhood adversity score, high (< 77) 39 (31.5) 24 (55.1) 2.27 (1.14 – 4.51) 0.019 Premorbid adjustment score, poor (≥ 4) 30 (24.2) 23 (48.9) 3.00 (1.48 – 7.07) 0.002 Acute psychosocial stressors, any 62 (50.0) 11 (23.4) 0.30 (0.14 – 0.65) 0.002 Duration of untreated continous psychosis, long (≥ 1 month) 34 (27.4) 26 (55.3) 3.27 (1.63 – 6.58) 0.001 Spontaneous dyskinesia, Schooler & Kane criteria 2 (1.9) 9 (24.3) 17.0 (3.48 – 83.3) <0.001 CGI-EI at index discharge, marked improvement 100 (80.6) 28 (59.6) 0.35 (0.17 – 0.73) 0.005 † Data are number (and percentages) of the stated features CGI-EI = Clinical global impression-Efficacy Index Conclusions FEP diagnoses other than schizophrenia or bipolar disorder should be considered as provisional. Considering baseline and background predictors of diagnostic change to schizophrenia may help to enhance diagnostic accuracy and guide therapeutic interventions. Disclosure of Interest None Declared
Donner et al. (2025) studied this question.